Fitting is typically performed to prepare a hearing aid for use. In this fitting, characteristics related to the hearing aid processing of the hearing aid (hearing aid characteristics) are set to suit the hearing ability of the hearing aid user. That is, first of all, the hearing ability of the user is measured over a range from low to high audible frequencies. Then, the hearing aid characteristics are adjusted on the basis of the hearing level data obtained by this measurement. More specifically, a value is set for hearing aid processing parameters that will determine the hearing aid processing to be executed by the hearing aid.
However, the work of adjusting a hearing aid often ends up taking a long time even for an experienced hearing aid adjuster.
Specifically, our perception of how we hear sounds varies from one person to the next, and this is also greatly affected by the measurement environment (such as the climate and the size of the measurement space), the time of day when the measurement is conducted, the physical condition of the hearing aid user, and so forth. These factors combine to make adjustment a time-consuming process. More specifically, once adjustment has proceeded a certain amount, the hearing aid adjuster gradually decides that the hearing aid processing parameters have been narrowed down to their final values that are suited to the hearing aid user.
Meanwhile, how a user hears sounds will sometimes vary compared to when the adjustment was started, due to external or internal factors. For instance, the user may conclude that a previous test sound was better than the current test sound. Therefore, the adjustment work frequently has to back up and the same work repeated, which tends to result in the adjustment taking a longer time.
In view of this, there have been attempts in the past at utilizing an interactive genetic algorithm (GA) in an effort to find an adjustment value semi-automatically, taking into account such subjective swings in evaluations of hearing (see Patent Literature 1, for example).